HIPPOCAMPUS 10:729–738 (2000) Cholinergic Basal Forebrain Is Critical for Social Transmission of Food Preferences Joanne Berger-Sweeney,1* Nancy A. Stearns,1 Karyn M. Frick,1 Binta Beard,1 and Mark G. Baxter2 1Department of Biological Sciences, Wellesley College, Wellesley, Massachusetts 2Department of Psychology, Harvard University, Cambridge, Massachusetts ABSTRACT: Studies using selective lesions of basal forebrain cholin- cholinergic decrease is associated with the magnitude ergic neurons suggest that these neurons play a role in attentional pro- of the memory impairment (Dunnett et al., 1987; cessing, but not learning and memory. However, the tests of learning and Fibiger, 1991). However, because the toxins used in memory used thus far have been restricted largely to spatial tasks. In the present study, we examined whether the cholinergic basal forebrain plays these earlier studies damage both cholinergic and non- a role in a form of nonspatial associative memory, the social transmission cholinergic neurons in the basal forebrain, it has been of food preferences. Sham-operated control rats were compared to rats difficult to assess the specific contributions of cholin- with 192 IgG-saporin lesions of the medial septum/diagonal band cholin- ergic neurons to memory processes. More recent stud- ergic projections to hippocampus or nucleus basalis magnocellularis/ substantia innominata cholinergic projections to neocortex. Both lesions ies using the selective cholinergic toxin 192 immuno- impaired 24-h retention of a learned social food preference relative to globulin G (IgG)-saporin to lesion the basal forebrain controls, despite performance on an immediate retention trial that was have yielded a different pattern of results from that of indistinguishable from controls. Moreover, 24-h retention of the socially earlier studies. 192 IgG-saporin lesions of the basal learned food preference correlated strongly with cholinergic enzymatic forebrain produce little or no impairments on spatial activity in the neocortex, but not in the hippocampus. Immunohistochem- ical data confirmed significant and selective lesion-induced cholinergic reference memory tasks (Berger-Sweeney et al., 1994b; depletions in the intended brain regions. These data provide evidence that Torres et al., 1994; Wenk et al., 1994; Baxter et al., the cholinergic basal forebrain, particularly the cholinergic projection to 1995) or on working memory tasks (Wenk et al., neocortex, is involved in the formation and/or retrieval of social memo- 1994; Walsh et al., 1995; Dornan et al., 1997; McMa- ries related to food preference, and suggest a role for cortical acetylcho- han et al., 1997; Chappell et al., 1998; Wrenn et al., line in consolidation of associative memory processes. Hippocampus 1999), despite depletions of cholinergic markers more 2000;10:729–738. © 2000 Wiley-Liss, Inc. severe than those produced by nonselective neurotox- KEY WORDS: medial septum; nucleus basalis; olfaction; 192 IgG-sa- ins. The general lack of impairments after selective porin; nonspatial learning and memory cholinergic lesions has led many researchers to ques- tion whether the cholinergic basal forebrain plays any role in learning and memory processes. A series of elegant studies from several laboratories INTRODUCTION suggests that the cholinergic system may play a role in selective attention rather than memory (Chiba et al., The role of the cholinergic basal forebrain in the formation or retrieval 1995; McGaughy et al., 1996; Baxter et al., 1997). Ex- of memories has been the subject of considerable debate recently (Chap- tensive depletions of neocortical acetylcholine from le- pell et al., 1998; Wrenn and Wiley, 1998). Studies using nonselective sions to the nucleus basalis magnocellularis/substantia neurotoxins such as ibotenic acid to lesion the basal forebrain report innominata (nBM/SI) result in robust impairments in lesion-induced impairments on a variety of tasks (Dekker et al., 1991; sustained attention (McGaughy et al., 1996) and in in- Gallagher, 1997). Furthermore, in some cases, the magnitude of the cremental attentional processing of conditioned stimuli (Chiba et al., 1995). In contrast, extensive depletion of hippocampal acetylcholine from lesions to the medial Grant sponsor: National Science Foundation; Grant number: 9976623; septum/vertical limb of the diagonal band of Broca (MS/ Grant sponsor: National Institute of Mental Health; Grant number: R03- VDB) impairs decremental attentional processing of con- MH60089. ditioned stimuli (Baxter et al., 1997). These studies point *Correspondence to: Joanne Berger-Sweeney, Department of Biological Science, Wellesley College, Wellesley, MA 02481. toward a role of the cholinergic basal forebrain in cogni- E-mail: [email protected] tive attentional processes, but not necessarily learning Accepted for publication 7 June 2000 and memory. © 2000 WILEY-LISS, INC. 730 BERGER-SWEENEY ET AL. Several studies, primarily using pharmacological manipulations, suggest that social learning and memory may be modulated by the MATERIALS AND METHODS central cholinergic system. In rodents, social recognition of a con- specific is a task dependent on short-term olfactory-related mem- Subjects ory processes (Ravel et al., 1994). Generally, drugs that reduce Thirty-eight male, 6-week-old Wistar rats were purchased from central cholinergic transmission impair the ability of an animal to Charles River, (Raleigh, NC). They were housed singly (4) or in recognize and remember previously encountered conspecifics, pairs (34) in a temperature- and humidity-controlled room on a whereas drugs that enhance cholinergic transmission improve 12/12-h light/dark schedule. Before pretraining for the social short-term social recognition memory in a dose-dependent fashion learning task began, the rats were allowed food (Harlan Teklad (Perio et al., 1989; Gheusi et al., 1994; Ravel et al., 1994; Winslow 22/5 Rodent Diet-W, Harlan Teklad, Madison, WI) and water ad and Camacho, 1995; Levy et al., 1997). Cholinergic stimulation of libitum. Surgery and behavioral testing were performed during the muscarinic receptors in the olfactory bulbs of rats appears to be light phase of the cycle. Rats were housed in an AAALAC Interna- necessary for chemosensory reception and the formation of short- tional accredited facility, and procedures were approved by the term olfactory memory (Ravel et al., 1994). Interestingly, the neu- Wellesley College Animal Care and Use Committee. ral processing of odor cues is synchronized to ongoing theta rhythms recorded in the hippocampus; cholinergic cells of the medial septum that project to hippocampus serve as pacemakers Surgery for theta rhythms (Bland, 1986). In total, these studies suggest that At 9 weeks of age, animals were randomly assigned to one of four the central cholinergic system is critically involved in processing of groups. The initial numbers of animals in each group were: MS/ odor cues and the formation and/or retrieval of social recognition VDB lesion, n ϭ 10; nBM/SI lesion, n ϭ 10; sham lesion to memories; however, the specific brain structures in which cholin- MS/VDB, nBM/SI, or intracerebroventricular n ϭ 10; no surgery, ergic modulation is critical have not been identified. n ϭ 8. Rats that received no surgery were designated as demon- In the current study, we sought to examine the role of the cho- strators, whereas the remaining rats were designated as observers. linergic basal forebrain system in a social learning and memory task Animals assigned to surgical groups were anesthetized with 80 that uses olfactory cues, but does not involve aversive stimuli or mg/kg ketamine and 5 mg/kg xylazine (i.m.), then placed in a spatial learning. This task, which examines a socially transmitted stereotaxic apparatus (Kopf Instruments, Tujunga, CA). Lesion food preference, exploits an ethologically meaningful behavior: an and sham surgeries were conducted by injection as described below animal’s ability to learn quickly and remember information per- using a motorized stereotaxic injector (Stoelting, Wood Dale, IL) taining to social olfactory cues (Galef and Wigmore, 1983; Strupp and 28-gauge needle Hamilton syringe filled with either 0.175 ␮ ␮ and Levitsky, 1984). In this task, an “observer” rat encounters g/ l 192 IgG-saporin in sterile phosphate-buffered saline (lesion another “demonstrator” rat that recently has eaten a food with a surgeries) or sterile phosphate-buffered saline (sham surgeries). distinctive scent. Then the observer rat is placed in a separate cage For MS/VDB surgery, two holes were drilled in the skull at ϭϩ and is given a chance to select food with the same scent smelled on stereotaxic coordinates anterior posterior (AP) 0.45 mm and ϭϮ the demonstrator’s breath or food with a different and unfamiliar medial lateral (ML) 0.6 mm from Bregma (Paxinos and Watson, 1998). Injections were made at two depths at each site, scent. Most observers will show a strong preference for the food ϭϪ Ϫ smelled on the demonstrator’s breath, presumably because this dorsal ventral (DV) 7.8 mm and 6.2 mm from the surface of the skull. Solutions were delivered at a rate of 0.05 ␮l/min, inject- indicates that the food consumed by the demonstrator was not ing a total of 0.3 ␮l at each of the DV ϭ -7.8 mm sites, and a harmful. This form of social learning involves the formation of a volume of 0.2 ␮l at each of the DV ϭϪ6.2 mm sites. The syringe specific stimulus-stimulus association in a single acquisition, with- was left in place for 6 min after each 0.3 ␮l injection and for 4 min out any primary reinforcement, and then the expression of the after each 0.2 ␮l injection to limit diffusion of solution into the memory in a different context (Bunsey and Eichenbaum, 1995). needle track. This form of social learning appears to be dependent on an intact For nBM/SI surgery, four holes were drilled in the skull at hippocampus (Winocur, 1990; Bunsey and Eichenbaum, 1995). stereotaxic coordinates AP ϭϪ0.75 mm and ML ϭϮ2.3mm Therefore, we hypothesized that basal forebrain lesions affecting (medial sites) and ML ϭϮ3.3 mm (lateral sites) from Bregma.